Journal of Experimental Marine Biology and Ecology, L 240 (1999) 37±52

The potential for suspension feeding bivalves to increase productivity

Bradley J. Petersona,* , Kenneth L. Heck Jr.b aDepartment of Marine Sciences, University of South Alabama, Mobile, AL 36688, USA bDauphin Island Sea Lab/University of South Alabama, P.O. Box 369, Dauphin Island, AL 36528, USA Received 15 January 1999; received in revised form 23 March 1999; accepted 30 March 1999

Abstract

Suspension feeding bivalves are commonly associated with seagrass habitats in the and . Biodeposits of some suspension feeding bivalves have been shown to be high in nitrogen and phosphorus. Consequently, ®lter feeding bivalves may act as a bentho- pelagic couple bringing planktonic production to the benthos, thereby elevating submerged aquatic vegetation growth by increasing the nutrients available to the rhizosphere. Laboratory feeding experiments were used to calculate the ®ltration rate of a typical suspension feeding bivalve americanus. Filtration rates were estimated to be 2.8760.82 l h21 g tissue dry weight21 . Consumption rates were estimated to be 9.4162.62 mg Chl a h21 g tissue dry weight21 . In addition, ®eld experiments were used to calculate mean biodeposition rates. Biodeposition rates were estimated to be 2.2560.36 g dry wt material g tissue dry weight day21 . Therefore, at mean ®eld densities M. americanus are capable of depositing 218 kg dry weight material m22 annually. These deposits will contain 215 g N and 7.1 g P. A ¯ower pot experiment demonstrated that the biodeposits of M. americanus were capable of increasing the pore water nutrient content and a mussel density manipulation in the ®eld revealed that the presence of mussels signi®cantly reduced leaf tissue C:N and C:P ratios. Pore water ammonium and phosphate concentrations were four times greater in the highest mussel density than in the control treatments and the lower leaf tissue C:N and C:P ratios in the presence of mussels established that this increased pore water nutrient was available to the seagrass, testudinum. Collectively, these experiments suggest that suspension feeding bivalves may be important resource conduits converting inaccessible PON and POP in the water column to elevated sediment nutrient levels within the rhizosphere available for absorption by submerged aquatic vegetation.  1999 Elsevier Science B.V. All rights reserved.

Keywords: Bentho-pelagic couple; Biodeposition; ± interactions; Seagrass; Thalassia testudinum; Suspension feeding bivalves; Modiolus americanus

*Corresponding author. Present address: Department of Biological Sciences, Florida International University, University Park, Miami, FL 33199, USA. Tel.: 11-305-348-1556; fax: 11-305-348-1986. E-mail address: petersob@®u.edu, (B.J. Peterson)

0022-0981/99/$ ± see front matter  1999 Elsevier Science B.V. All rights reserved. PII: S0022-0981(99)00040-4 38 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52

1. Introduction

Dense assemblages of ®lter feeding bivalves (e.g. mussels) remove suspended matter from the water column and deposit it as feces or pseudofeces on the bottom. Biodeposits from ®lter feeding bivalves may signi®cantly contribute to the total suspended load in shallow coastal environments (Haven and Morales-Almo, 1966; Tenore and Dunstan, 1973; Kraeuter, 1976; Tsuchiya, 1980). Such biologically mediated sedimentation has the capacity to greatly exceed passive physical processes in the deposition of ®ne sediments in estuaries and coastal environments (Biggs and Howell, 1984). These biodeposits represent a potentially signi®cant energy source to consumers (Vahl, 1980; Kautsky, 1981; Newell et al., 1982; Stewart, 1987). For example, Newell et al. (1982) calculated that one third of the annual production of particulate matter in a kelp bed area was from fecal matter originating from suspension feeding benthic . After bacterial enrichment, a signi®cant proportion was reingested, leading these authors to conclude that a mussel `fecal loop' plays a signi®cant role in the circulation of nutrients and organic matter in a kelp community. Additionally, ®lter feeders transform suspended material by changing particle size distributions in the water column and by converting particulate material into dissolved constituents or biomass via metabolism (Dame et al., 1980). Dame et al. (1985) suggested that feeding by bivalve aggregates may act as a positive feedback loop in which particulate nitrogen (phytoplankton) consumed by the ®lter feeding bivalves is rapidly remineralized to ammonium (NH4 ). This ammonium is then available for plant growth. This model has been envisioned as a nitrogen retention mechanism as well as a process which accelerates the nitrogen cycle (Dame et al., 1989). Previous studies have illustrated that ®lter feeding bivalves may control phytoplankton abundance through feeding and nutrient excretion activities (Dame et al., 1980, 1985, 1989, 1991; Cloern, 1982; Of®cer et al., 1982; Carlson et al., 1984; Prosch and McLachlan, 1984; Nichols, 1985; Doering and Oviatt, 1986; Boucher and Boucher- Rodoni, 1988; Dame and Dankers, 1988; Yamamuro and Koike, 1993). These studies generated great interest in the role that ®lter feeding bivalves have on phytoplankton growth dynamics and biomass. However, the in¯uence that suspension feeding bivalves have on submerged aquatic vegetation has been largely ignored. The potential for suspension feeding bivalves to affect the growth dynamics of marine angiosperms may be great. Seagrass productivity is limited primarily by nutrient and light availability. If both are equally important, this creates an apparent environmental incongruity for . Increasing water column nutrient levels results in elevated plankton and epiphytic growth which may decrease light availability for seagrasses. Therefore, seagrasses are usually limited to areas with relatively low water column nutrient concentrations. But, unlike phytoplankton, which rely exclusively on water column nutrient sources, seagrasses primarily takes up nutrients from the sediments by using roots (Agami and Waisel, 1986). Previous studies have demonstrated that sediment pore water is the primary source of nutrients for seagrass growth (Stewart, 1987). Because biodeposits of some suspension feeding bivalves are high in nitrogen and phosphorus (Kautsky and Evans, 1987, Jaramillo et al., 1992), ®lter feeding bivalves can potentially transfer planktonic production from the water column to the benthos via feces B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 39 and pseudofeces, and thereby enhance submerged aquatic vegetation growth by increasing the nutrients available in the rhizosphere. Bertness (1984) investigated the interaction of cordgrass, Spartina alterni¯ora, and the semi-infaunal mussel, Geukensia demissa, demonstrating that the presence of this mussel at densities as high as 900 individuals m22 increased net production and the above ground±below ground biomass ratio of S. alterni¯ora. Recently, Reusch et al. (1994) suggested that the blue mussel, Mytilus edulis, fertilizes eelgrass, Zostera marina, growth by the deposition of feces and pseudofeces. However, the effect that biodeposits of ®lter feeding bivalves have on seagrass productivity remains uncertain. One conspicuous plant±animal association within seagrass habitats of the Gulf of Mexico and Caribbean Sea involves turtle grass, Thalassia testudinum, and the semi- infaunal suspension feeding tulip mussel, Modiolus americanus (Leach) (Rodriguez, 1959; Jackson, 1973; Young and Young, 1982; Lyons, 1989; Valentine and Heck, 1993). In St. Joseph Bay, Florida, extensive monotypic stands of T. testudinum contain patchily distributed clusters of M. americanus, making this location ideal for experimental manipulation and thus for determination of the signi®cance of a seagrass±bivalve interaction. Using both laboratory and ®eld experiments, the following questions were addressed: (1) what are the particle consumption, ®ltration and biodeposition rates of Modiolus americanus? (2) how does the nutrient content of biodeposits compare to that of naturally sedimenting material? (3) can the addition of biodeposits increase the nutrient content of sediment pore water? and (4) are these nutrients available to Thalassia testudium?

2. Study site

The potential effects of mussels on seagrass assemblages were observed in St. Joseph Bay, Florida, in the northeastern Gulf of Mexico (308 009 N, 858 309 W) during the summer (May±August) of 1996. St. Joseph Bay is a protected shallow coastal embayment where salinities usually range from 30 to 36½ (Stewart and Gorsline, 1962; Folger, 1972; this study). Temperatures vary seasonally from approximately 8 to 308C (this study), and the mean tidal range is 0.5 m (Rudloe, 1985). The bay is oligotrophic with water column nitrogen and phosphorus values seldom exceeding 3 and 0.2 mM, respectively (J. Pennock, unpublished data). Phytoplankton abundance is also low, usually below 5mg/l. Therefore, photosynthetically active radiation (PAR) is high, with approximately 40% of measured light at the water surface reaching the seagrass canopy (Heck and Valentine, in press). St. Joseph Bay supports an extensive seagrass habitat occupying ¯ 26 km2 of shallow bay bottom (McNulty et al., 1972). This seagrass habitat is dominated by large monospeci®c stands of Thalassia testudinum interspersed with smaller patches of , unvegetated sand ¯ats, and small patches of Syringodium ®liforme (Iverson and Bittaker, 1986). Seagrass production is highly seasonal with blade biomass and density peaking near 150 g AFDW m22 and 3000 blades m22 , respectively, during summer months (Iverson and Bittaker, 1986). Only the shallowest portions of the 40 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 seagrass habitat are exposed during low tides, and wave energy is minimal. Associated with these T. testudinum beds is the semi-infaunal tulip mussel, Modiolus americanus, which is found in densities as high as 2000 individuals m22 (Valentine and Heck, 1993). Mean densities of M. americanus at the study site were 625 individuals m22 . Due to the great variability in mussel size between individuals, a linear regression of shell length to g tissue dry weight was constructed for M. americanus:

tissue dry wt. weight 520.474 1 0.018*shell length (1)

(r 2 50.92; n5328). This allowed a more precise comparison to be made on g dry wt tissue m22 rather than on individuals m22 . Based on this regression at mean ®eld densities, there was 426 g tissue dry wt in each m2 of sediment surface.

3. Materials and methods

3.1. Flow-through laboratory experiment

Consumption rate (matter ingested per unit time per unit mussel weight) (Hildreth and Crisp, 1976) was estimated using ¯ow-through techniques. Using a modi®ed equation of

Northby (1976) consumption at time t (Ct ) can be estimated by the equation:

Cttt5 f(Qb 2 Qm ) (2)

where f is the ¯ow rate, Qbt is the concentration of the rate-indicator substance at the out¯ow of the control tank and Qmt is the concentration of the rate-indicator substance at the out¯ow of the experimental tank (Frechette and Bourget, 1985). The use of this equation requires the following assumptions: (1) that the control tanks are mounted in parallel with the experimental tanks, (2) there is zero consumption in the control tank, (3) that the ¯ows between the experimental and control tanks are equal, and (4) that the experimental organisms occupy a negligible proportion of total tank volume. The rate-indicator substance for this ¯ow-through experiment was an algal mono- culture of the genus Thalassiosira. A randomized block design was used in this experiment. Six replicate ¯ow-through chambers were mounted in parallel and randomly assigned to either treatment (three control; three mussels present). The bottom of each chamber was covered by approximately 3 cm of agar. Six mussels (approximate length 45 mm; constituting approximately 2% of the tank volume) were placed vertically into the agar of the experimental chambers and allowed to acclimate for 30 min. The ¯ow rates in all six chambers were synchronized prior to each of the consumption rate estimations. The experiment was repeated four times with mussels that had not been used in previous trials (n512 for each treatment). After the initiation of the experiment, a known volume of water (250 ml) was collected from the out¯ow of each chamber every 5 min for 30 min. Output concentrations from control and experimental ®ltration chambers were measured by nephelometry, using a Turner Associates Model 111 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 41

¯uorometer. In addition, acetone extracted chlorophyll a was collected on Whatman GF/C glass ®ber ®lters used as an index of phytoplankton biomass was determined by ¯uorometry (Strickland and Parsons, 1972).

3.2. Biodeposition tubes ®eld experiment

In the ®eld, a randomized block design was used to estimate biodeposition rate and test for differences in nutrient content of sedimenting material. Four replicate pairs of PVC cylinders (50 cm in length, 19 cm ID) were tied to racks and attached to the sediment surface perpendicular to the dominant tidal ¯ow. The design of the biodeposi- tion tube followed that of Kautsky and Evans (1987) and Jaramillo et al. (1992). One cylinder of each replicate pair was randomly assigned to the experimental treatment (n54 for each treatment). The tops of the cylinders were covered by VexarE mesh (20 mm ID), and a vexar mesh shelf was placed 50 mm below the cylinder lids. Mussels were collected 24 h prior to the initiation of the experiment and held in aquaria allowing them to void their digestive tracts. Shell morphometrics were recorded and each individual was marked with numbered electrician wire tape. At the initiation of the experiment, mussels were randomly selected and placed between the mesh in the experimental cylinder of each pair. To avoid size speci®c biodeposition rates, mussels of all size ranges were randomly selected. Mussel biomasses within the cylinders ranged between 236 and 680 g tissue dry weight m22 , which corresponded with the average ®eld mussel biomass occurring in St. Joseph Bay (426 g tissue dry weight m22 ). Natural sedimenting material was collected in all the cylinders, while biodeposits were collected only in those containing mussels. Biodeposition was calculated as the amount of material collected in each mussel cylinder minus the average sedimentation obtained in the control cylinders. After 27 days the cylinders were retrieved from the ®eld. The collected material within the cylinders were transferred to buckets and suspended while ten 40-ml aliquots were removed from each cylinder's sediment sample. Each aliquot was drawn through a glass ®ber ®lter (2.4 cm Whatman GF/C glass ®ber ®lters) and dried at 608C to a constant mass. All ten ®lters were averaged together to estimate the total suspended load for each cylinder. Five ®lters were then randomly assigned for nutrient analysis or for ashing at 5008C for 5 h. Total carbon and nitrogen of the sedimented material was determined by combustion with the Carlo-Erba NA1500 (Sharp, 1974). Particulate organic phosphorus (POP) was converted to inorganic phosphorus by high temperature combustion. Residue polyphosphates were hydrolyzed with addition of hydrochloric acid. Total phosphorus was measured by the reactive phosphorus method. Extinction rates were read by spectrophotometric analysis at 885 nm (Fourqurean et al., 1992a). Differences between control and experimental cylinders in C:N and C:P ratios of the sedimented material were tested with one-way analysis of Variance (ANOVA).

3.3. pot ®eld experiment

A completely randomized design was used to conduct a 65-day in situ manipulation experiment within a monospeci®c grassbed of Thalassia testudinum in St. Joseph Bay, 42 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52

Florida. Twelve plastic containers (15.24 cm in diameter; 0.75 l volume) were ®lled with sediment from which the organic material had been digested with 10% H22 0 . The containers were randomly assigned to one of three mussel density treatments (0 individuals m22 , 500 individuals m22 , and 1000 individuals m22 ) and placed ¯ush with the sediment surface of the grassbed. A single sediment plug (10 cc) was extracted from each container at the initiation of the experiment to establish initial nutrient con- centrations and at the conclusion of the experiment to test for an increase in porewater nutrient levels. Pore water was collected by centrifugation and analyzed on an Alpkem Rapid Flow Analyzer 2 (RFA/2). Differences between the three treatments in sediment porewater ammonium and phosphate were tested with a one-way ANOVA.

3.4. Leaf tissue nutrient content experiment

A completely randomized design was used to perform density manipulations of live mussels in nine randomly assigned 0.25 m2 plots within a heavily vegetated portion of a turtle grass meadow. Plot margins were marked by anchoring PVC frames to the sediment. Possible translocation of stored nutrients through the seagrass out of/or into the plots was prevented by severing the rhizomes around the perimeter of each plot. This experiment was conducted for three months (March 1996±May 1996). Mussels were added randomly to plots at treatment densities of 0, 500, and 1500 individuals m22 (n53 for each density). Within 7 days of planting, the mussels had reattached themselves into natural positions. These densities approximate the range of abundances most often observed in the bay (Valentine and Heck, 1993). To document any potential changes in nitrogen and phosphorus concentrations in blades due to treatment effects, ®ve shoots were randomly collected from each plot after 3 months and biomass speci®c changes in the concentrations of carbon, nitrogen and phosphorus in the dried blades were measured following Fourqurean et al. (1992a). Leaves were gently scraped and washed in ¯owing tap water to remove epibionts and sediments that had adhered to the leaves. These washed samples were dried to a constant mass and homogenized by milling to a ®ne powder. The elemental contents of C, N and P of these seagrass leaves were then ascertained for each treatment. Leaf tissue C and N were determined by oxidation in a Carlo Erba Model 1500 CNS analyzer. Phosphorus content was measured using a modi®cation of the method presented in Solorzano and Sharp (1980) by Fourqurean et al. (1992b) for total particulate phosphorus determi- nation.

4. Results

4.1. Flow-through laboratory experiment

Out¯ow concentrations from the control chambers were 4.2860.61 mg Chl a (n512) while those of the experimental chambers were 3.4860.61mg Chl a (n512). Using the modi®ed equation of Northby (1976), the consumption rate of Modiolus americanus was B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 43 calculated to be 9.4162.62 mg Chl a h21 g21 tissue dry wt., while the ®ltration rate was calculated to be 2.8760.82 l h21 g21 tissue dry wt.

4.2. Biodeposition tubes ®eld experiment

Natural sedimentation within the biodeposition tubes ranged from 33 to 173 g dry wt. m22 day21 , while biodeposition varied from 431 to 1171 g dry wt. m22 day21 (equivalent to 1.013±2.75 g dry wt. g tissue dry wt.21 day21 ). The corresponding values for ashfree dry wt. were 26±112 g m22 day21 and 234±677 g dry wt. m22 day21 (0.55±1.59 g dry wt. g tissue dry wt.21 day21 ) for sedimentation and biodeposition, respectively. Combining all replicates, the average natural sedimentation rate was 120 g dry wt m22 day21 and the average biodeposition rates was 959 g dry wt m22 day21 (2.25 g dry wt. g tissue dry wt21 day21 ). C:N ratios were signi®cantly lower for biodeposits than for naturally sedimented material ( p,0.001, F522.29, df57) (Fig. 1). Correspondingly, C:P ratios were signi®cantly lower for biodeposits ( p50.03, F55.77, df57). However, N:P ratios were not signi®cantly different between the two treatments ( p50.277, F51.32, df57).

4.3. Flower pot ®eld experiment

Pore water nutrient concentrations increased dramatically from the initiation to the conclusion of the experiment. The mean values for ammonium (NH4 ) in the control, 500 mussel m22 and 1000 mussel m22 treatments at the initiation of the experiment were

151, 608 and 227 mM NH4 . After 65 days in the ®eld, the pore water concentration of NH4 in the treatments were 1391, 5224 and 8803 mM NH4 (Fig. 2a). At the highest mussel density treatment, pore water ammonium levels increased by a factor of 38. 22 Similarly, the mean values for phosphate (PO4 ) in the control, 500 mussel m and 1000 mussel m22 treatments at the initiation of the experiment were and 101, 56 and 63 mM PO4 and at the conclusion of the experiment the pore water concentrations of PO4 were 65, 185 and 461 mM (Fig. 2b). Over the course of the experiment, pore water PO4 levels decreased in the control treatments and increased in the highest mussel treatments by a factor of 7. The pore water nutrient concentrations of both NH44 and PO demonstrated a signi®cant positive response to the presence of ®lter feeding bivalves ( p50.001, F519.73, df511 and p,0.001, F523.19, df511, respectively).

4.4. Leaf tissue nutrient content experiment

Thalassia testudinum leaf tissue C:N exhibited a signi®cant decline from 16.3160.38 in the 0 mussel treatment, to 14.8560.67 in the 500 mussel treatments, to 13.3760.38 in the 1500 mussel treatments ( p50.001, F526.306, df52) (Fig. 3). Differences were signi®cant for all treatment combinations (Student±Newman±Keuls method). A similar pattern was observed for treatment effects on leaf tissue C:P which declined from 774.62625.84 to 702.52639.80 to 658.74628.21 with increasing mussel densities ( p50.012, F510.113, df52). However, leaf tissue N:P demonstrated no signi®cant change with increasing mussel densities ( p50.62, F50.51, df52). 44 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52

Fig. 1. (a) C:N, (b) N:P and (c) C:P ratios of Modiolus americanus biodeposits and naturally sedimenting material from the biodeposition tube experiment (Bars5mean61 S.D.). B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 45

Fig. 2. (a) Pore water ammonium concentration comparison between control, 500 mussels m22 and 1000 mussels m22 at day 0 and day 65, respectively. (b) Pore water phosphate concentration comparison between control, 500 mussels m22 and 1000 mussels m22 at day 0 and day 65, respectively (Bars5mean61 S.D.; differing letters indicate signi®cant differences between treatments using Student±Newman±Keuls multiple comparison method). 46 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52

Fig. 3. (a) Leaf tissue C:N ratio comparison between between control, 500 mussels m22 and 1500 mussels m22 . (b) Leaf tissue C:P ratio comparison between control, 500 mussels m22 and 1500 mussels m22 (Bars5mean61 S.D.; differing letters indicate signi®cant differences between treatments using Student± Newman±Keuls multiple comparison method). B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 47

5. Discussion

The experimental consumption rates of Modiolus americanus for this study were compared to ®ve models used in the literature to estimate consumption of suspended particles by ®lter feeding bivalves (Table 1). Models from the literature are all functions of size, temperature, or both. Four of the models predicted ®ltration rates within a close margin of the actual measured ®ltration rates for M. americanus. However, the model of Cloern (1982) signi®cantly overestimated the actual ®ltration rate (Table 1). In these laboratory experiments, M. americanus was demonstrated to be capable of consuming twice the amount of Chl a present in the water column in St. Joseph Bay. It is possible, however, that the low Chl a content of the water column in St. Joseph Bay may be the direct consequence of the high consumption rates of M. americanus and other suspension feeding bivalves. Organic matter deposited as feces may represent a signi®cant proportion of nutrient potentially available to submerged aquatic vegetation. C:N and C:P ratios have been employed to assess the nutritional values of food, and detritus (Russell-Hunter, 1970; Kautsky and Evans, 1987; Parson et al., 1977). Lower C:N and C:P ratios indicate higher concentrations of nitrogen or phosphorus. Accordingly, comparisons of C:N ratios between naturally sedimenting particles and mussel biodeposits yield insights into the characteristics of both materials. Biodeposits of M. americanus were greatly enriched in both nitrogen and phosphorus, but C:N ratios for biodeposits analyzed in this study were higher than that of previously published studies (Jordan and Valiela, 1982; Kautsky and Evans, 1987; Jaramillo et al., 1992). C:N and C:P ratios were still signi®cantly lower for biodeposits than that of naturally sedimenting material. Approximately half of the particulate nitrogen and carbon consumed by mussels is expelled as feces (Jordon and Valiela, 1982; Hawkins and Bayne, 1985), while the corresponding amount for phosphorus may be as high as 94% (Kuenzler, 1961). Bivalve molluscs can eliminate between 5 and 50% of total nitrogen excretion in the form of amino acids, but this phenomenon is highly seasonal and is usually associated with periods of starvation and/or gametogenesis when catabolism of protein provides a reserved energy supply (Bayne and Scullard, 1977). Previous studies have shown that ammonia excretion rates for bivalve molluscs increase with temperature and show a positive logarithmic relationship to body weight (Duerr, 1968; Bayne and Scullard, 1977).

Table 1 Comparison of ®ltration models from the literature with the estimates from the ¯ow-through experiments for Modiolus americanus (L5shell length, T5temperature, W5weight) Filtration model Filtration estimate Reference (ml ind21 min21 ) [(L 0.96)(T 0.95)]/2.95 24.73 Doering and Oviatt (1986) 5.12L 0.967 21.92 Doering and Oviatt (1986) 2.59W 0.73 23.15 Coughlan and Ansell (1964) 0.76W 20.40 17.82 Of®cer et al. (1982) 168W 0.67 65.86 Cloern (1982) Actual 20.38 This study 48 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52

Table 2 Comparison of biodeposition rates from the literature with the estimates from the biodeposition tubes for Modiolus americanus Biodeposition rate Reference (g N day21 g tissue dry wt21 ) Mytilus edulis 5.05531024 Kautsky and Evans (1987) Choromytilus chorus 2.06631023 Jaramillo et al. (1992) Mytilus chilensis 2.37831023 Jaramillo et al. (1992) Modiolus demissus 1.20031023 Jordan and Valiela (1982) Modiolus americanus 2.4731023 This study 6.9231023 This study

Biodeposition rates of nitrogen for Modiolus americanus are consistent with those previously reported for most other suspension feeding bivalves (Table 2). The estimates of this ®eld experiment reveal that if all of the nutrient within the biodeposits of M. americanus were available to Thalassia testudinum, then on an annual basis biodeposits would provide eight times greater nitrogen and seven times greater phosphorus than required for maximal leaf growth in T. testudinum (Patriquin, 1972). If all of this nutrient were directly available to the plant, then the requirements for maximal leaf growth in Thalassia would be achieved through the biodeposits of 1/6th of the biomass of M. americanus present in the average m2 of sediment surface in St. Joseph Bay. The C:N:P ratios of have been used to assess the nutrient status of phytoplankton (Red®eld, 1958) and macrophytes (Gerloff and Krombholtz, 1966). The amount of nitrogen or phosphorus, relative to carbon, in plant tissues is a function of the availability of N or P in the environment. Forqurean et al. (1992b) found that leaf tissue C:N and C:P of T. testudinum decreased with increasing pore water soluble reactive phosphorus and ammonium. Furthermore, the N:P ratio of seagrass leaf tissue re¯ected the relative availability of N and P in the environment. In this study, pore water nutrients increased dramatically in the presence of M. americanus and a subsequent decline in leaf tissue C:N and C:P was observed with increasing mussel densities. This indicates that the increased nutrient concentration of the sediments is biologically available to the plant and that plants in association with suspension feeding bivalves have an increased nutrient content within their leaf tissue. Thus, T. testudinum has more resources available for growth when associated with M. americanus. Numerous organisms have been shown to have a profound in¯uence on the communities they live in through habitat or resource modi®cation (Jones et al., 1994). Previous investigators have demonstrated a positive effect of suspension feeding bivalves on plant production. Bertness (1984) found that the fecal material from Geukenia demissa stimulated growth of Spartina alterni¯ora on which the mussels were attached. Similarly, Reusch et al. (1994) documented enhanced growth of Zostera marina in the presence of the suspension feeding blue mussel, Mytilus edulis. These studies emphasize the stimulating effect of biodeposition on benthic plant production. In the grass system of St. Joeseph Bay, the mussels can be envisioned as transforming the unavailable particulate organic nitrogen and phosphorus in the water column into B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 49

Fig. 4. Conceptual model of the role of Modiolus americanus in seagrass assemblages of Thalassia testudinum. The annual estimates of ®ltration and biodeposition are listed.

accessible nitrogen and phosphorus in the sediments (Fig. 4). The estimates of M. americanus biodeposition rates suggest that these organisms are capable of transferring a signi®cant amount of nitrogen and phosphorus from the water column to the benthos. Consequently, their ecological impact may be great. It is conceivable that M. ameri- canus, by increasing the sediment nutrient level, may create new habitable areas for colonization by T. testudinum or maintain suf®cient nutrient levels for the continued existence of T. testudinum in stressful environments.

Acknowledgements

Funding for this project was provided in part by the Mississippi±Alabama Sea Grant, University of South Alabama and the Dauphin Island Sea Lab. We thank the Dauphin Island Sea Laboratory faculty and staff for their help in all phases of this project. J. Valentine, J. Cowan, M. Bertness and three anonymous reviewers provided helpful comments on various drafts. MESC Contribution No. 306 and SERC contribution No. 99. 50 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52

References

Agami, M., Waisel, Y., 1986. The ecophysiology of roots of submerged vascular plants. Physiologiw veg. 24, 607±624. Bayne, B.L., Scullard, C., 1977. Rates of nitrogen excretion by species of Mytilus edulis (: ). J. Mar. Biol. Assoc. UK 57, 355±369. Bertness, M.D., 1984. Ribbed mussels and Spartina alterni¯ora production in a New England salt marsh. Ecology 65, 1794±1807. Biggs, R.B., Howell, B.A., 1984. The estuary as a sediment trap: alternative approaches to estimating its ®ltering ef®ciency. In: Kennedy, V.S. (Ed.), Proc. 7th Biennial Int. Estuarine Res. Conf., Virginia Beach, Virginia, October 23±26, 1983, The Estuary as A Filter, pp. 107±129. Boucher, G., Boucher-Rodoni, R., 1988. In situ measurement of respiratory metabolism and nitrogen ¯uxes at the interface of oyster beds. Mar. Ecol. Prog. Ser. 44, 229±238. Carlson, D.J., Townsend, D.W., Hilyare, A.L., Eaton, J.F., 1984. Effect of an intertidal mud ¯at on plankton of the overlying water column. Can. J. Fish. Aquat. Sci. 41, 1523±1528. Cloern, J.E., 1982. Does the benthos control phytoplankton biomass in South San Francisco Bay? Mar. Ecol. Prog. Ser. 9, 191±202. Coughlan, J., Ansell, A.D., 1964. A direct method for determining the pumping rate of siphonate bivalves. J. Conseil 29, 205±213. Dame, R.F., Dankers, N., 1988. Uptake and release of materials by a Wadden Sea mussel bed. J. Exp. Mar. Biol. Ecol. 118, 207±216. Dame, R.F., Zingmark, R., Stevenson, H., Nelson, D., 1980. Filter feeding coupling between the estuarine water column and benthic subsystems. In: Kennedy, V.S. (Ed.), Estuarine Perspectives, Academic Press, New York, pp. 521±526. Dame, R.F., Wolaver, T.G., Libes, S.M., 1985. The summer uptake and release of nitrogen by an intertidal oyster reef. Neth. J. Sea Res. 19, 265±268. Dame, R.F., Spurrier, J.D., Wolaver, T.G., 1989. Carbon, nitrogen, and phosphorus processing by an oyster reef. Mar. Ecol. Prog. Ser. 54, 249±256. Dame, R.F., Dankers, N., Prins, T., Jongsma, H., Smaal, A., 1991. The in¯uence of mussel beds on nutrients in the Western Wadden Sea and Eastern Scheldt estuaries. Estuaries 14, 130±138. Doering, P.H., Oviatt, C.A., 1986. Application of ®ltration rate models to ®eld populations of bivalves: an assessment using experimental mesocosms. Mar. Ecol. Prog. Ser. 31, 265±275. Duerr, F.G., 1968. Excretion of ammonia and urea in some species of molluscan prosobranch snails. Comp. Biochem. Physiol. 26, 1051±1059. Folger, D.W., 1972. Characteristics of estuarine sediments of the United States. US Geol. Surv. Prof. Paper 742, 1±95. Fourqurean, J.W., Zieman, J.C., Powell, G.V.N., 1992. Phosphorus limitation of primary production in Florida Bay: evidence from C:N:P ratios of the dominant seagrass. Limnol. Oceanogr. 37, 162±171. Fourqurean, J.W., Zieman, J.C., Powell, G.V.N., 1992. Relationships between porewater nutrients and seagrasses in a subtropical carbonate environment. Mar. Biol. 114, 57±65. Frechette, M., Bourget, E., 1985. Energy ¯ow between the pelagic and benthic zones: factors controlling particulate organic matter available to an intertidal mussel bed. Can. J. Fish. Aquat. Sci. 42, 1158±1165. Gerloff, G.C., Krombholtz, P.H., 1966. Tissue analysis as a measure of nutrient availability for the growth of angiosperm aquatic plants. Limnol. Oceanogr. 11, 529±537. Haven, D., Morales-Almo, R., 1966. Aspects of biodeposition by oysters and other ®lter feeders. Limnol. Oceanogr. 11, 487±498. Hawkins, A.J.S., Bayne, B.L., 1985. Seasonal variation in the relative utilization of carbon and nitrogen by the mussel Mytilus edulis: budgets, conversion ef®ciencies and maintenance requirements. Mar. Ecol. Prog. Ser. 25, 181±188. Hildreth, D.I., Crisp, D.J., 1976. A corrected formula for calculations of ®ltration rate of bivalve molluscs in an experimental ¯owing system. J. Mar. Biol. Assoc. UK 56, 111±120. Iverson, R.L., Bittaker, H.F., 1986. Seagrass distribution and abundance in the eastern Gulf of Mexico waters. Estuar. Coast. Shelf Sci. 22, 577±602. B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52 51

Jackson, J.B.C., 1973. The ecology of molluscs of Thalassia communities. Jamaica, West Indies. I. Distribution, environmental physiology, and ecology of common shallow-water species. Bull. Mar. Sci. 23, 313±350. Jaramillo, E., Bertran, C., Bravo, A., 1992. Mussel biodeposition in an estuary in southern Chile. Mar. Ecol. Prog. Ser. 82, 85±94. Jones, C.G., Lawton, J.H., Shachak, M., 1994. Organisms as ecosystem engineers. OIKOS 69, 373±386. Jordan, J.E., Valiela, I., 1982. A nitrogen budget of the ribbed mussel, Geukensia demissa, and its signi®cance to nitrogen ¯ow in a New England salt marsh. Limnol. Oceanogr. 27, 75±90. Kautsky, N., 1981. On the trophic role of the blue mussel (Mytilus edulis L.) in the Baltic coastal ecosystem and the fate of the organic matter produced by the mussels. Kiel. Meeres. (Sonderh.) 5, 454±461. Kautsky, N., Evans, S., 1987. Role of biodeposition by Mytilus edulis in the circulation of matter and nutrients in a Baltic coastal ecosystem. Mar. Ecol. Prog. Ser. 38, 201±212. Kraeuter, J.N., 1976. Biodeposition by salt-marsh . Mar. Biol. 35, 215±223. Kuenzler, E.J., 1961. Phosphorus budget of a mussel population. Limnol. Oceanogr. 6, 400±415. Lyons, W.G., 1989. Nearshore marine ecology at Hutchinson Island, Florida: 1971±1974. XI. Mollusks. Florida Mar. Res. Publ 47. McNulty, J.K., Lindahl, W.C., Sykes, J.E., 1972. Cooperative Gulf of Mexico estuarine inventory and study, Florida: phase I, area description. NOAA Tech. Rep. NMFS Circ (368). Newell, R.C., Field, J.G., Grif®ths, C.L., 1982. Energy balance and signi®cance of microorganisms in a kelp bed community. Mar. Ecol. Prog. Ser. 8, 103±113. Nichols, F., 1985. Increased benthic grazing: an alternative explanation for low phytoplankton biomass in Northern San Francisco Bay during the 1976±1977 drought. Estuar. Coast. Shelf Sci. 21, 379±388. Northby, J.A., 1976. A comment on rate measurements in open systems. Limnol. Oceanogr. 21, 180±182. Of®cer, C.B., Smayda, T.J., Mann, R., 1982. Benthic ®lter feeding: a natural eutrophication control. Mar. Ecol. Prog. Ser. 9, 203±210. Parson, T.R., Takahashi, M., Hargrave, B. (Eds.), 1977. Biological Oceanographic Processes, 2nd ed, Pergamon Press, Oxford. Patriquin, D.G., 1972. The origin of nitrogen and phosphorus for growth of the marine angiosperm Thalassia testudinum. Mar. Biol. 15, 35±46. Prosch, R.M., McLachlan, A., 1984. The regeneration of surf-zone nutrient by the sand mussel, Donax serra Roeding. J. Exp. Mar. Biol. Ecol. 80, 221±233. Red®eld, A.C., 1958. The biological control of chemical factors in the environment. Am. Sci. 46, 561±600. Reusch, T.B.H., Chapman, A.R.O., Groger,È J.P., 1994. Blue mussels Mytilus edulis do not interfere with eelgrass Zostera marina but fertilize shoot growth through biodeposition. Mar. Ecol. Prog. Ser. 108, 265±282. Rodriguez, G., 1959. The marine communities of Margarita Island, Venezuela. Bull. Mar. Sci. Gulf Caribb. 9, 237±280. Rudloe, A., 1985. Variation in the expression of lunar and tidal rhythms in the horseshoe , Limulus polyphemus. Bull. Mar. Sci. 36, 388±395. Russell-Hunter, W.D., 1970. Aquatic Productivity, Macmillan, New York. Sharp, J.H., 1974. Improved analysis for `particulate' organic carbon and nitrogen from seawater. Limnol. Oceanogr. 19, 984±989. Solorzano, L., Sharp, J.H., 1980. Determination of total dissolved phosphorus and particulate phosphorus in natural waters. Limnol. Oceanogr. 25, 754±760. Stewart, A.J., 1987. Responses of stream to grazing minnows and nutrients: a ®eld test of interactions. Oecologia 72, 1±7. Stewart, R.A., Gorsline, D.S., 1962. Recent sedimentary history of St. Joseph's Bay, Florida. Sedimentology 1, 256±286. Strickland, J.D.H., Parsons, T.R. (Eds.), 1972. A Practical Handbook of Seawater Analysis, Fish. Res. Board Can, pp. 49±52. Tenore, K.R., Dunstan, W.M., 1973. Comparison of feeding and biodeposition of three bivalves at different food levels. Mar. Biol. 21, 190±195. Tsuchiya, M., 1980. Biodeposit production by the mussel Mytilus edulis L. on rocky shores. J. Exp. Mar. Biol. Ecol. 47, 203±222. 52 B.J. Peterson, K.L. Heck / J. Exp. Mar. Biol. Ecol. 240 (1999) 37 ±52

Vahl, O., 1980. Volume of water pumped and particulate matter deposited by the Iceland scallop (Chlamys islandica) in Balsfjord, Northern Norway. In: Freeland, H.J., Farmer, D.M., Levings, C.D. (Eds.), Fjord Oceanography, Plenum Press, New York, pp. 639±643. Valentine, J.F., Heck, Jr. K.L., 1993. Mussels in seagrass meadows: their in¯uence on macroinvertebrate abundance and secondary production in the Northern Gulf of Mexico. Mar. Ecol. Prog. Ser. 96, 63±74. Yamamuro, M., Koike, I., 1993. Nitrogen metabolism of the ®lter-feeding bivalve Corbicula japonica and its signi®cance in primary production of a brackish lake in Japan. Limnol. Oceanogr. 38, 997±1007. Young, D.K., Young, M.W., 1982. Macrobenthic invertebrates in bare sand and seagrass (Thalassia testudinum) at Carrie Bow Cay, Belize. In: Rutzler, K., MacIntyre, I.G. (Eds.), The Atlantic Barrier Reef Ecosystem at Carrie Bow Cay, Belize. I. Structure and Communities, Smithsonian Press, Washington, DC, pp. 115±126.